Andreas Aardal Hanssen wrote:
For backwards compatibility, we could allow selects on subfolders of INBOX
by translating these into our "actual" format, so when selecting INBOX.a
you're actually selecting "a". LIST would return the new format, and
perhaps clients would "detect" that its subscribed folders no longer
"exist" and then unsubscribe from them automatically. I'd have to test
this, but I'm optimistic.

Because of Binc's current behavior, we would always have to allow a select
on INBOX.a to translate into a select on "a".
I think this is bad. A select on "INBOX.a" should be a select on "INBOX.a". A select on "a" should be a select on "a".

At the same time, we could only allow subscription to folders using the
new format, and only allow creates with the same rule. Thereby moving the
users to a format where no subfolders are allowed under INBOX unless they
are there "already".
Is there a reason to prevent folders under INBOX? The standard doesn't seem to limit such things. If so, why? If not, then I say leave them.

I suppose my question is this: Is bincimap currently translating selects on "a" to "INBOX.a"? If so, this behavior should be deprecated. If not, then simply allowing a means to create and maintain root folders--and designating them, perhaps without a leading '.' would be the best solution. Subfolders of Maildir/ that are detected as Maildirs are offered up as either subfolders of INBOX if the beign with '.' or as their own top-level folders if they do not.

Anyway, it's a good thing to do things like this in early stage of
development and not postpone it until the server gets really popular. I think we can make this work - I'm eager to try.
I agree, changes like these should be made early on. I'm eager to move bincimap into my production servers.

As a side note, I think a solution for root folders that uses a directory other than Maildir/ is a bad idea. Some of us operate mail systems where the only user specific directory is their Maildir/, they don't have a 'home' directory as they are not actual users on the system. (Obviously their Maildir/ is not named Maildir/)

C=)

--
--------------------------------------------------------------------------
Better the hard truth than the comforting fantasy. -- Carl Sagan
--------------------------------------------------------------------------
Caskey <caskey*technocage.com> /// TechnoCage Inc.
--------------------------------------------------------------------------
A presumption on your part does not constitute an obligation on my part.

Reply via email to